Participatory Accountability and Collective Action◊
Abigail Barr Truman Packard Danila Serra
University of Nottingham The World Bank Florida State University
5 March 2012
Abstract
There is general agreement that the existence of participatory institutions is a necessary condition for accountability, especially where top-down institutions are malfunctioning or missing. In education, the evidence on the effectiveness of participatory accountability is mixed. We argue that participation is a social dilemma and therefore depends, at least partly, on individuals’ propensity to cooperate with others for the common good. This being the case, the mixed evidence could be owing to society-level heterogeneities in individuals’ willingness and ability to overcome collective action problems. We investigate whether individuals’ propensity to cooperate plays a role in parents’ decisions to participate in both a school accountability system – i.e., a “short route” to accountability – and parliamentary elections – i.e., a “long route” to accountability, – by combining survey data on 1800 individuals’ participation decisions with measures of their willingness to contribute to a public good in the context of a very simple, clearly defined laboratory experiment. We conduct our study in a new democracy, Albania, involving parents of children enrolled in primary schools. Our findings confirm that, both across individuals within communities and across communities, the decision to hold teachers and school directors accountable directly through participation at the school level, and indirectly through political participation correlates with cooperativeness in a simple public goods game.
JEL CODES: C93, D72, H4, O12 KEYWORDS: accountability; participation; collective action; public good game
◊ We thank Klaus Abbink, Margaret McConnell, Ragan Petrie, John Ryan, Paula Cordero Salas, Tim Salmon, Sara Solnick, Pedro Vicente, Jane Zhang and seminar participants at Florida State University, Southern Methodist University, the University of Texas Dallas, Wesleyan University, the 2011 SEA conference, and the 2011 North-American ESA meeting for useful comments and suggestions. We are indebted to the Institute for Development Research Alternatives (IDRA), who collected the data, and especially Auron Pasha, Enkelejda Pashaj and Adela Gjergjani for their technical excellence and professionalism. The findings, interpretations and conclusions expressed herein are those of the authors and do not necessarily reflect the view of the World Bank Group, its Board of Directors or the governments they represent. The usual disclaimers apply. Corresponding author. Address: Department of Economics, Florida State University, 258 Bellamy, Tallahassee, FL 32306. Email: [email protected].
1
1. Introduction
Accountability refers to “the act of holding public officials and service providers answerable for processes and outcomes, and imposing penalties if specified outputs and outcomes are not delivered” (Lewis and Petterson, 2009). Effective accountability requires the establishment of clear rules and responsibilities, monitoring, and actual enforcement of predetermined penalties when rules are not followed and responsibilities are not met. While governments have a central role in setting the rules and responsibilities of public service providers through formal institutional arrangements, monitoring and penalty-enforcement cannot be supplied by the government alone.1 Problems of conflict of interest and possible collusion between the monitors and the monitored2 call for a more participatory approach to the problem of accountability, i.e. an approach that directly involves service recipients (see World Bank, 2004). Under the participatory approach, governments put institutions in place that allow citizens to play an active role in keeping service providers accountable. There are two possible routes to participatory accountability: a long route and a short route (World Bank, 2004). The long route connects recipients of the services and policymakers; in the ideal institutional setting, with well-functioning electoral systems, free from patronage and clientelism, and with perfect information on the standards that public service provision should meet, service recipients hold policymakers accountable for the delivery of public services through the exercise of the vote. However, the ideal setting is hardly ever a reality, especially in less developed countries, where the route between service recipients and policymakers is itself often long and tortuous. Decentralization may shorten the route to a certain extent, but electoral inefficiencies, elite capture and imperfect information may still hinder accountability. 3 In such contexts, the shortest “local” route to accountability, connecting service recipients directly to service providers, may be more viable (World Bank, 2004). We argue that, although the type and amount of participation required in the short and long routes is different, in both cases an individual’s decision to participate has the characteristics of a collective action problem, i.e., a situation where individuals “face choices in which the maximization of the short-term self-interest yields outcomes leaving all participants worse off than feasible alternatives” (Ostrom, 1998). Individual participation in holding service providers directly to account or in voting involves a private cost. However, individual participation benefits others, as efficacy requires a
1 In the words of Paul Collier, “accountability is about restraints on government power and so depends upon the institutions conceded by government under pressure from citizens. Unlike other public goods, accountability cannot be provided by the government alone” (http://bostonreview.net/BR34.4/collier.php). 2 For instance, in Kenya, Kremer and Chen (2001) reported that when headmasters were charged with the duty of monitoring and reporting primary school teacher attendance so that bicycles could be awarded to those with good attendance records, the headmasters indicated that all their teachers had sufficiently good records to deserve a bike, while independent verification revealed otherwise. 3 See Oates (1972) and Azfar et al. (2001) for general discussion of the possible limitations of decentralization.
2 significant number of participants; hence, participation has the characteristics of a social dilemma and is likely to be subject to “free-rider” problems (Samuelson, 1954; Olson, 1965; Grossman and Hart, 1980). In education, reforms aiming to improve accountability by increasing parental participation in schools – also called school-based management reforms – have been and are currently being conducted in a number of places around the world. 4 The few scientific evaluations of their effectiveness have focused on two issues: 1) the institution of participatory mechanisms; and 2) the extent to which parents are aware of such mechanisms. Existing studies (Banerjee et al, 2010; Blimpo and Evans, 2011; Duflo et al., 2011; Pradhan et al. 2011) provide mixed evidence of success,5 yet the reasons for the contrasting results obtained in different societies are unclear.6 We argue that variations in the success of participatory interventions could be owing, at least partly, to variations in individuals’ willingness to act cooperatively in collective action problems. 7 Experimental studies employing public goods games have shown that individuals vary markedly in their willingness to cooperate with others for the common good (Isaac and Walker, 1988; Fehr and Gachter, 2000; Fischbacher et al., 2001; Bowles and Gintis, 2002; Fischbacker and Gachter, 2010, among others).8 Moreover, a few additional experimental studies (Heinrich et al., 2005; Herrmann et al., 2008; Gachter et al., 2012) have highlighted that a significant proportion of the variation in individuals’ willingness to cooperate occurs at the society level.9 We investigate whether, both at the individual- and the community-level, willingness to cooperate with others for the common good is associated with willingness to participate in both the short- and the long-route to accountability. We conduct our study in Albania, a country that slowly transitioned
4 In practice, school-based management (SBM) programs may or may not give decisional power to the parents. Some programs give parents decisional power over the school budget only; others involve parents also in the choice of the curriculum and textbooks; others give parents responsibilities in the hiring and firing of teachers; others invite parents to participate in the development of a school improvement plan, and possibly in fund-raising, and so on and so forth. School- based management programs may also exclude parents from participating in the decision process. See Wohlstetter and Odden (1992) for a categorization of school-based management models, and Barrera-Osorio et al. (2009) for a comprehensive review of school-based management reforms recently undertaken all over the world. 5 For mixed evidence not relating to the education sector, see Olken (2007) and Bjorkman and Svensson (2008). 6 Investigations based on observational rather than experimental data – also called “retrospective studies” – conducted in the US (Houtenville and Conway, 2007) in Argentina (Eskeland and Filmer, 2007), in El Salvador (Jimenez and Sawada, 1999), in Mexico (Gertler et al., 2008; and Skoufias and Shapiro, 2006), among others, all show a significant impact of participatory programs and parental involvement on education outcomes. 7 This is an important issue that has received little attention in the existing empirical literature. As stated by Platteau (2008), in order for participation to be effective, “community members must be able to use the available information jointly in a way that creates some action, that is, they must be able to come together, share and discuss their knowledge and be ready to act on it” (p. 128). 8 For comprehensive reviews of the experimental literature relating to behavior in public goods experiments see Ledyard (1995) and Chaudhuri (2011). 9 One important society characteristic that has been shown to be negatively correlated with individual contributions to a public good (Habyarimana et al. 2007), participation in community activities (Alesina and La Ferrara, 2000), community- based monitoring of public service providers (Bjorkman and Svensson, 2010), and local public good provision (Miguel and Gugerty, 2004) is ethnic fragmentation.
3 from communism to a parliamentary democracy in the 1990s, and that, like other new democracies in Eastern Europe and the Balkans, is characterized by the lack of an active civil society, owing to policies implemented by former regimes to eliminate groups mediating between the individual and the state or control these groups in order to prevent organized opposition to the regime (Lipset, 1993).10 Our analysis is based on a comprehensive survey of 1800 parents and 900 teachers, randomly selected from 180 nationally representative primary schools. Isolating the effect of willingness to cooperate with others in collective action scenarios is empirically challenging. We generate a direct, experimental measure of individuals’ propensities to cooperate with others in social dilemmas, by involving the parents and teachers in a simplified version of the public goods game first introduced by Cardenas et al. (2009). We then correlate cooperativeness in the game with parents’ decision to participate in both the short route to accountability, i.e. the decision to participate in the election of parent class representatives, and the long route to accountability, i.e., the decision to vote in the most recent parliamentary elections. The nature of the survey sampling also allows us to investigate whether and how the distribution of willingness to cooperate varies spatially within Albania. We found that parents who behaved cooperatively in the public goods game are significantly more likely to have participated in the election of parent class representatives who go on to elect parent representatives to the school board. This relationship remains significant after controlling for the degree to which parents are informed about aspects of local accountability in schools. We also found that parents who behaved cooperatively in the public good game are more likely to (claim that they) have voted in the latest national elections. Our data show significant variation in both participation and cooperativeness across districts. Our district-level analysis confirms that districts characterized by a higher proportion of cooperative parents have a higher degree of participation in school accountability and a higher turnout in the 2009 elections according to official records. Our results suggest that one important reason why, in some societies more than in others, creating participatory institutions and providing information about the existence and functioning of these institutions might fail to induce participation and bottom-up accountability is that potential participants might be unwilling or unable to overcome collective action problems. Our study also contributes to the literature on the external validity of public goods laboratory experiments by providing evidence of a correlation between behavior within a lab-type public goods game and behavior in naturally occurring decision-contexts within which inter-personal cooperation is also salient. While the public goods game has been shown to correlate with fishermen’s productivity when pooling their catches with other fishermen (Carpenter and Seki, 2011), and with their propensity (not) to over-exploit common fishing grounds (Fehr and Leibbrandt, 2011), this is, to the best of our knowledge, the first study providing
10 For instance, in Albania religious practices were officially banned in 1967, to be reinstated in the 1990s.
4 evidence of a correlation between individuals’ behavior in the public goods experiment and their willingness to cooperate with others in participatory accountability systems and civic engagement. The paper is arranged as follows. Section 2 describes the Albanian context focusing, in particular, on the education sector. Section 3 describes the school survey and the public goods experiment. Section 4 presents descriptive statistics about the surveyed parents and explores individual and community characteristics affecting parents’ cooperativeness in the public goods game. Section 5 reports our empirical results, and Section 6 provides some conclusions.
2. The Albanian context
2.1 A new democracy
Albania was the last country in Europe to participate in the “third wave of democracy” (Huntington, 1991). Despite the high levels of economic growth experienced during the democratic transition, Albania remains one of the poorest in Europe, with a per capita GDP of 8,000 USD (2010 international PPP $). Albania’s relative underdevelopment has been attributed, in part, to the slowness of its transition to democracy relative to other Eastern European countries. The new constitution, which established that the “sovereignty in the Republic of Albania belongs to the people” (article 2) and that “governance is based on a system of elections that are free, equal, general and periodic” (article 1), was adopted in 1998. Since then, there have been three parliamentary elections – not without controversies concerning electoral fraud, protests and boycotts by the losing party. In each, voter turnout has been around 50%. Albania’s slow transition to democracy has been attributed to the unique characteristics of its communist regime: its isolation from other countries, including the Soviet bloc, for half a century; the elimination of intellectuals (including western-educated Albanians); the abolition of religious practices; and the harsh persecution of opponents to the regime. With the end of communism, a new era began. According to Elbasani (2004), the recently acquired freedom was interpreted and understood by Albanians as the “unhindered pursuit of personal gains at the expense of society and public good”. One of the consequences of the repression of civic organizations during the communist regime is the lack of an active civil society, i.e. formal and informal organizations (with or without political objectives) and voluntary community participation, in the newly formed democracy of Albania.11 This is consistent with the evidence that totalitarian regimes destroy social capital, however
11 According to Talifi (2008), most of the efforts to build a civil society in Albania – primarily undertaken by donor-based NGOs – have relied on awareness campaigns and capacity building, and have emphasized the importance of an informed society and electoral base. The inherent shortcoming of this approach is the assumption that by informing people about
5 conceptualized, i.e., as the density of individuals’ informal networks, the extent of interpersonal trust, the density of voluntary organizations or the extent to which individuals play the cooperative solution in collective action problems (see Paldan and Svendsen, 2000; Rose, 1993; Smolar, 1996).
2.2. Spatial differences Twelve counties and thirty-six districts constitute the territory of Albania. For historical, geographical and cultural reasons the north and the south of the country are substantially different. Historically, northern Albania has been more isolated than the south, relying on a patriarchal tribal system oriented around the institutions of clan and blood feuds. Southern rural communities have had more market-oriented institutions such as agricultural labor markets, because of the historical presence of latifundia (large estates) and the hiring of migrating labor. Moreover, the southern region has been culturally more open to change and to the influence of Western values. (Doll, 2003; Shala, 1997; Lawson and Saltmarshe, 2000).12 The North-South cultural division also reflects a strong political polarization: the Northern regions are traditionally and historically allied to the Democratic Party (in office since 2005) while the Southern regions support the Socialist Party (in office from 1997 to 2005).13 As shown empirically by Gërxhani and Schram (2000, 2009), Albanian politics is dominated by clientelism and patronage, whereby the party in power favors its own supporting regions over the rest of the country, and voters support their “own” party no matter what. Besides the north-south distinction, Albania can be further divided in four main geographical areas – Coastal, Centre, Mountain and Tirana – which significantly differ in terms of economy, infrastructure and level of development. The Mountain region, in the north-east of the country, is most remote and poorest; the economy relies mainly on heavy industry, owing to the presence of minerals. Besides Tirana, the most developed areas are found on the coast, where both farming and trade opportunities abound. The central areas rely on light industries (food processing, wood, clothes etc.) and partly on agriculture. Migration from the Mountain region to the Coast and Tirana has increased markedly over the last 15 years, and explains the increasing heterogeneity in the coastal population. Figure 1A in the Appendix shows the north-south divide and the four main geographical areas in Albania.
democratic and participatory institutions, participation will follow. Talifi (2008) argues that “this approach of civil society has decreased rather than increased public participation in the process, because simply telling people to participate is not a good enough approach to contribute to the democratization of the country”. 12 The North and the South also differ in the main dialects spoken, i.e. the Gheg in the North and the Tosk in the South. 13 Since 1997, Albania relies on a proportional voting system. The Democratic Party (DP) and the Socialist Party (SP) are the two dominant parties. In the 2009 elections, the DP obtained 40.18% of votes, all from the Northern regions, and the PD obtained 40.85% of votes, all from the Southern regions.
6
2.3 The education sector
In the years following the transition, there was a sharp decline in the coverage and quality of social services. School enrolment rates dropped and, in some areas, have still not recovered to pre-transition levels. Although 99% of the adult population is literate, the quality of education is low, as shown by Albania’s performance in the 2009 Program for International Student Assessment (PISA). More than 50% of 15 years old students found it difficult “to use continuous texts unless the texts are short and clearly sign-posted; and even with such texts, they are unlikely to be able to do more than identify a main idea or find explicitly stated information.”14 The Ministry of Education and Science (MoES) is the central government body responsible for implementing education policies and managing the education system. This responsibility is exercised by staff in the MoES and in twelve administration entities (REDs) functioning at the county level. The MoES is directly responsible for the development of curricula, the selection of textbooks, the structure of the academic year, setting the pay and workloads of teachers, the allocation of resources among local education institutions, and teacher training (World Bank, 2005). The 12 REDs are responsible for the delivery of primary and secondary education in their respective districts, but they are not decentralized government units. They are responsible for the appointment and transfer of teachers, the distribution of administrative materials to schools, inspections of schools, and planning and supervising the construction of new facilities and the rehabilitation of existing ones. However, the RED directors are appointed by the MoES and have no authority with respect to the amount of funding centrally allocated to their region. The institutional framework currently in place relies on a system that assigns the task of monitoring teachers and imposing penalties for under-performance to administrative units that are not directly held to account by service recipients, i.e., it relies on top-down accountability. 15 However, the MoES has recently been promoting community engagement in school governance structures through parent committees and school boards. Parent committees are composed of elected parent class representatives, and school boards are composed of two or more (depending on the size of the school) parent representatives, one teacher representative, a student representative, a community representative and a RED representative. Parent class representatives are elected by the parents of pupils in their respective classes at the beginning of the school year. They, then, elect representatives from amongst themselves to the school boards.
14 See http://www.portugal.gov.pt/pt/GC18/Documentos/ME/PISA_2009_1.pdf 15 The existing top-down system seems poor in terms of accountability is functioning poorly. According to our school survey, it is very common for teachers to believe that there are no negative repercussions to bad performance: about 50% indicated that they would not receive a penalty if they underperformed. Moreover, those who thought they would be penalized were unclear with respect to what kind of penalty they would receive.
7
While school boards have always existed, they have recently been given greater responsibilities for school governance. The main functions of the School Board are to examine and approve (by voting) the school’s mid-term and annual plans and the school’s annual financial report. The board also has decisional power for the adoption of curricula and textbooks, as well as the school’s budget relating to contributions from the community or other donors, or revenues from school activities. School Board members may also discuss problems relating to the performance of teachers, or school directors, following complaints from parents, students, or teachers. Since 2008, the school director is not a member of the school board and, although he/she can still participate in meetings, he/she cannot vote.
- Figure 1 here -
Figure 1 depicts the accountability relationships in (pre-university) education in Albania. Participatory accountability can take the “long route”, which connects parents directly to the ruling government and only indirectly to the school, through voting in parliamentary elections. Participatory accountability can also take a shorter route, connecting parents directly to the school through the election of parent class representatives that, in turn, elect parent members to school boards.
3. The School Stakeholder Survey
We conducted the Albania School Stakeholder Survey as part of the World Bank’s Accountability for Better Governance Program. The survey was primarily aimed at investigating the role that parents play in the institutional framework within which primary schools currently operate.
3.1 Survey design and implementation
The data collection took place between October and December 2009 in a representative sample of 180 primary schools, offering grades one to nine. The sampling strategy relied on stratification at the district level; the number of schools randomly selected to participate in the survey in each district depended on the number of pupils attending primary school in that district.16 For each school, three
16 The sample was drawn from the list of all public primary schools in Albania. There are 2691 public basic level schools in Albania, of which 12 percent are located in urban areas. Out of the 2691 primary schools, we defined as “eligible” those schools with five or more pupils in grade three, and ten or more pupils in grade seven. Only 1623 of the 2691 primary schools in the complete list met this eligibility criterion. However, only 9 percent of all Albanian public primary school
8 and seven students were randomly selected from grades three and six respectively. One teacher of the third graders and four teachers of the seventh graders were randomly selected to participate in the survey together with the parents of the selected pupils. Therefore, in each school we collected data from ten parents and five teachers, leading to total sample sizes of 1800 parents and 900 teachers.17 Besides demographics, we collected information about parents’ involvement in the school accountability system, i.e. whether they participated in the latest elections of parent class representatives. We also elicited data on the degree to which each parent was informed about the existence of local accountability institutions, and their involvement in other aspects of their children’s education, such as the number of meetings with the head teacher in the previous semester and how often they helped their children with homework. In designing the School Stakeholder Survey, we purposely replicated some relevant LSMS questions concerning education and social capital, with the aim to be able to check the extent to which our sample of parents is representative of the Albanian adult population. With respect to social capital, we collected information about: 1) membership in voluntary organizations; 2) participation in any community activity in the past year; 3) number of relatives and close friends among the surveyed parents (a measure of social ties in the community); 4) beliefs about community members’ willingness to cooperate with each other in case of a water shortage; 5) a measure of generalized trust. As part of the survey, all parents and teachers were also involved in three behavioral games. In this paper we focus only on parents’ behavior in a dichotomous public goods game.18 After the games, we collected information about the existing social ties between the experimental participants.
3.2 The public goods experiment
In order to generate a direct measure of individuals’ propensities to coordinate and cooperate with others to solve collective action problems, we conducted a public goods experiment. Laboratory experiments facilitate the measurement of individuals’ values, beliefs and preferences that cannot be captured in survey data. Experimentally generated measures of individual preferences for cooperation in the context of a public goods game have been shown to correlate with individuals’ behavior in pupils attend ineligible schools. The list of eligible schools was then divided into 36 strata according to district and either six, four or two schools was randomly sampled from each, the number depending on the number of pupils attending primary school in the district. Since the district of Tirana contained twice the percentage of pupils as well as twice the number of independent eligible schools, it was divided into two strata, Tirana Municipality and Tirana District, leading to a total of 37 strata. In all but two counties, the majority of surveyed schools where located in villages or hamlets with less than 3000 inhabitants. The average number of pupils per school is 327, although this number falls to 200 if we exclude the urban schools. 17 The MoES assisted in the implementation of the survey by demanding the full cooperation of school directors and teachers in the sampled schools. 18 For a full description of the Albania School Stakeholder survey and behavioral experiments, see Serra, Barr and Packard (2011).
9 natural life in a number of contexts. Fehr and Leibbrandt (2011) found that Brazilian fishermen who behaved more selfishly in a public goods game were more likely to over-exploit common fishing grounds. Carpenter and Seki (2011) found that Japanese fishermen’s behavior in a public goods game predicted their productivity when pooling their catches with other fishermen.19 Public goods games have also been used following randomized interventions to identify changes in individuals’ attitudes and preferences. For instance, Fearon, Humphrey and Weinstein (2011) employed a public goods game to assess the impact of a community driven development initiative in Liberia on individuals’ ability to overcome collective action problems. Similarly, Attanasio et al. (2009) employed a public goods game to compare individuals’ propensities to cooperate with each other in two communities in Colombia, one that had received conditional cash transfers for over two years, and one that had not. We adapted the binary public goods game of Cardenas et al. (2009) to the Albanian context. All surveyed parents and teachers participated in the experiment on their school premises. Thus, each experimental session included 15 subjects. Each participant was given a voucher and had to decide whether to invest their voucher either in a group account or a private account. If an individual invested in the group account, he/she would get 100 LEK plus 100 LEK multiplied by the number of other participants investing in the group account:20